A highly selective aerobic oxidation process catalyzed by electron-deficient nitroarenes via single electron transfer processes

Article abstract of DOI:10.1021/jo025916y

A versatile and simple method for aerobic oxidation of various aromatics containing an oxygen-functionalized benzylic carbon is reported. Results from oxidation experiments with methyl aryl ketones, benzaldehydes, benzylic alcohols, and mandelic acid are reported; all provide high yields of the corresponding aromatic carboxylic acids. By means of response surface methodology and mechanistic interpretation it is revealed that the oxidation process operates catalytically using electron-deficient nitroarenes, such as 1,3-dinitrobenzene, 1,2,3-trifluoro-4-nitrobenzene, and 2,4-difluoro-1-nitrobenzene, with molecular oxygen as the terminal oxidant. The reaction is carried out in a basic solution of potassium tert-butoxide in tert-butyl alcohol. Since the terminal oxidant is molecular oxygen and only 5-10 mol % of, e.g., 1,3-dinitrobenzene as catalyst is used, the new method represents an environmentally benign alternative to, for example, the well-known haloform reaction. The method is also a convincing alternative when transition metal free conditions are required.

Full text of DOI:10.1021/jo025916y

Evaluation Only. Created with Aspose.PDF. Copyright 2002-2021 Aspose Pty Ltd.
A High ly Selective Aer obic Oxid a tion P r ocess Ca ta lyzed by
Electr on -Deficien t Nitr oa r en es via Sin gle Electr on Tr a n sfer
P r ocesses
Hans-Rene´ Bjørsvik,* Lucia Liguori, and J ose´ Angel Vedia Merinero
Department of Chemistry, University of Bergen, Alle´gaten 41, N-5007 Bergen, Norway
hans.bjorsvik@kj.uib.no
Received May 8, 2002
A versatile and simple method for aerobic oxidation of various aromatics containing an oxygen-
functionalized benzylic carbon is reported. Results from oxidation experiments with methyl aryl
ketones, benzaldehydes, benzylic alcohols, and mandelic acid are reported; all provide high yields
of the corresponding aromatic carboxylic acids. By means of response surface methodology and
mechanistic interpretation it is revealed that the oxidation process operates catalytically using
electron-deficient nitroarenes, such as 1,3-dinitrobenzene, 1,2,3-trifluoro-4-nitrobenzene, and 2,4-
difluoro-1-nitrobenzene, with molecular oxygen as the terminal oxidant. The reaction is carried
out in a basic solution of potassium tert-butoxide in tert-butyl alcohol. Since the terminal oxidant
is molecular oxygen and only 5-10 mol % of, e.g., 1,3-dinitrobenzene as catalyst is used, the new
method represents an environmentally benign alternative to, for example, the well-known haloform
reaction. The method is also a convincing alternative when transition metal free conditions are
required.
In tr od u ction
processes as results of these investigations: (1) using
manganese and cobalt nitrates as catalysts with air or
oxygen as the terminal oxidant,6 and (2) using 1,3-
dinitrobenzene as a stoichiometric oxidant for obtaining
carboxylic acids from methyl aryl ketones benzaldehydes
and benzylic alcohols.7 The latter method was developed
on the basis of the alkaline nitrobenzene oxidation
method,8-10 which has been widely used in oxidation and
structure elucidation studies of lignin. The original
nitrobenzene oxidation method requires rather drastic
reaction conditions using a very alkaline environment
and elevated temperature (180-190 °C) and pressure
(10-12 atm.). Our recently reported method,7 which
utilizes the more powerful oxidant 1,3-dinitrobenzene,
requires considerably less drastic conditions. A reaction
temperature of only 100 °C at atmospheric pressure is
now required, but a strong alkaline medium is still
crucial for the process.
Functionalized aromatic carboxylic acids boast a vast
application area ranging from monomers for bulk poly-
mers and special polymers to fine chemicals and key
intermediates for the synthesis of numerous pharmaceu-
tical chemicals. 3,4-Dimethoxybenzoic acid (veratric acid)
is such a compound and is used in the synthesis of
mebeverine,1 vesnarinone,2 and itopride,3 which are
antispasmodic cardiotonic and gastric prokinetic agents,
respectively. Acetovanillone, industrially obtained as a
side product from the lignosulfonate oxidation process for
production of vanillin,4 is used as industrial feedstock for
veratric acid. Veratric acid is produced industrially from
acetovanillone through a simple two-step process consti-
tuted by a methylation and an oxidation step. The latter
utilizes the haloform reaction as recently improved and
optimized by us.5 As a part of this process improvement
and development project we initiated several other
investigations of more basic character to look for other
oxidation methods that hopefully could be converted into
benign and environmentally friendly industrial scale
processes. We have recently reported two new oxidation
Nevertheless, an obvious drawback exists even with
our modified alkaline nitrobenzene oxidation process; the
fact that the process requires the oxidant in stoichiomet-
ric quantities implies production of reduction products
of 1,3-dinitrobenzene.
On the basis of analogous single electron transfer
(SET) processes such as in the stoichiometric 1,3-dini-
trobenzene oxidation method,7 we have recently com-
* To whom correspondence should be addressed. Fax: +47 55 58
94 90.
(1) Kralt, T.; Moed, H. O.; Lindner, A.; Asma, W. J . (N.V. Philips’
Gloeilampenfabrieken). Patent DE 1126889; Chem. Abstr. 1963, 58,
27206.
(2) Yang, Y. H.; Tominaga, M.; Nakagawa, K.; Ogawa, H. (Otsuka
Pharmaceutical Co., Ltd., J apan). Patent BE 890942; Chem. Abstr.
1982, 97, 72386.
(6) Minisci, F.; Recupero, F.; Fontana, F.; Bjørsvik, H.-R.; Liguori,
L. Synlett 2002, 610.
(7) Bjørsvik, H.-R.; Liguori, L.; Minisci, F. Org. Process Res. Dev.
2001, 5, 136.
(3) Itoh, Y.; Kato, H.; Koshinaka, E.; Ogawa, N.; Nishino, H.;
Sakaguchi, J . (Hokuriku Pharmaceutical Co., Ltd., J apan). Patent EP
0 306 827; Chem. Abstr. 1989, 111, 153371.
(4) Bjørsvik, H.-R.; Minisci, F. Org. Process Res. Devel. 1999, 3, 330.
(5) Bjørsvik, H.-R.; Norman, K. Org. Process Res. Dev. 1999, 3, 341.
(8) Freudenberg, K.; Lautsch, W.; Engler K. Ber. Dtsch. Chem. Ges.
1940, 73, 167.
(9) Lautsch, W.; Plankenhorn, E.; Klink, F. Angew. Chem. 1942, 53,
108.
(10) Leopold, B. Acta Chem. Scand. 1950, 4, 1523.
10.1021/jo025916y CCC: $22.00 © 2002 American Chemical Society
Published on Web 09/20/2002
J . Org. Chem. 2002, 67, 7493-7500
7493

Evaluation Only. Created with Aspose.PDF. Copyright 2002-2021 Aspose Pty Ltd.
Bjørsvik et al.
SCHEME 1
municated11 a new catalytic oxidation process that we
believed to be made up by a base-catalyzed autoxidation
path and a pathway catalytic in 1,3-dinitrobenzene with
molecular oxygen as the terminal oxidant. This paper
presents a full account of this new oxidation process
providing improved conditions as well as new insights
into the mechanism of the highly selective and high
yielding catalytic oxidation process. The oxidation process
is still based on molecular oxygen as the terminal oxidant
with an electron-deficient nitroarene, e.g., 1,3-dinitroben-
zene, as mediator for the oxidation of aromatics contain-
ing oxygen-functionalized benzylic carbon.
reaction between the nitroarene radical anion 4 and the
carbon-centered radical 5 leads to formation of the labile
anion 6. The succeeding step entails fission of the weak
N-O bond as indicated in Scheme 1. This bond fission
gives nitrosoarene 7 as the first reduction product and
2-oxo-2-phenylethanoic anion 8 as the partly oxidized
product. Nitrosoarene 7 and 2-oxo-2-phenylethanoic an-
ion 8 may be formed via a “cage recombination” of the
radicals 4 and 5. This is also a highly reasonable
mechanism since the solvent and the reaction tempera-
ture have been shown to be two very important factors
in enabling the process to operate.
Further oxidation of the anion 8 gives the ultimate
oxidation product 9 according to the reaction pathway
8 f 12 f 13 f 14 f 9 of Scheme 1 that is analogous to
the steps 1 f ... f 8. The mechanism in Scheme 1
provides a framework that we have used to develop a new
catalytic oxidation cycle using molecular oxygen as the
terminal oxidant.
Meth od s a n d Resu lts
Russell and co-workers12 and Ayscough and co-work-
ers13 have reported conclusive evidence that carbanions
react with nitroarenes by SET processes with formation
of nitroarene radical anions and carbon-centered radicals.
On the basis of these studies we believe that the intrinsic
principle of the 1,3-dinitrobenzene oxidation process we
have recently reported7 relies on an initial SET step. The
mechanism of this process is clearly complex; our pro-
posal of a mechanism for the SET process is reproduced
with several extensions in Scheme 1. The SET step is
based on a carbanion or alkoxide ion 2 that carries out a
single electron transfer to the electron-deficient aromatic
ring of 1,3-dinitrobenzene 3. The reaction yields a nitro-
group-centered radical anion 4 and a carbon- or oxygen-
centered radical in the substrate 5. During our explora-
tion of this reaction we have discovered that it is essential
to use a substantial quantity of a strong base in order to
produce a sufficient amount of the enolate anion 2 from
the methyl aryl ketone 1. In the following step a coupling
In an endeavor to develop this oxidation process into
a catalytic process we have investigated several methods
that could hopefully oxidize the different reduction
products, viz., 1-nitro-3-nitrosobenzene 7, N-(3-nitrophen-
yl)hydroxylamine 10, 3-nitrophenylamine 11, etc., back
to 1,3-dinitrobenzene 3. Scheme 2 summarizes two such
systems that were tried for the in situ oxidation of these
compounds. The first system tried used transition metal
salt (Cu,2+ Co,2+ or Mn2+)-catalyzed oxidation of the
reduction products of 1,3-dinitrobenzene with molecular
oxygen as the terminal oxidant. The yields achieved are
reported in Scheme 2 and correspond, although only
roughly, to the quantity of 1,3-dinitrobenzene used as
catalyst. Thus it was concluded that the only oxidation
mechanism operating was the stoichiometric one as
described in Scheme 1 or autoxidation as described later
in this section. Investigations using transition metal salts
as cocatalyst with molecular oxygen as the terminal
oxidant were therefore abandoned.
(11) Bjørsvik, H.-R.; Liguori, L.; Rodr´ıguez Gonza´lez, R.; Vedia
Merinero, J . A. Tetrahedron Lett. 2002, 43, 4985.
(12) Russell, G. A.; J anzen, E. G. J . Am. Chem. Soc. 1962, 84, 4153.
(13) Ayscough, P. B.; Sargent, F. P.; Wilson, R. J . Chem. Soc. 1963,
5418.
7494 J . Org. Chem., Vol. 67, No. 21, 2002

Evaluation Only. Created with Aspose.PDF. Copyright 2002-2021 Aspose Pty Ltd.
Highly Selective Aerobic Oxidation Process
SCHEME 2
SCHEME 4
SCHEME 3
radical anions 17 were oxidized to yield two nitrobenzene
radical anions 19. On the other hand, if the reaction was
performed without oxygen present, the nitrosobenzene
radical anions 17 dimerized to N,N′-diphenyldiazene
N-oxide 20.
However, analogous catalytic oxidation using 1,3-
dinitrobenzene as the working mediator was further
investigated.
McKillop and co-workers14-16 have demonstrated per-
borates and percarbonates as superb oxidants for the
oxidation of anilines and derivatives into the correspond-
ing nitroderivatives. Using these oxidants, we have set
up a comparable catalytic oxidative scheme using either
sodium perborate or sodium percarbonate as the terminal
oxidant with 1,3-dintrobenzene as the working oxidant
for the methyl aryl ketone. However, only traces or very
modest yields were obtained from these experiments. In
earlier studies of the nitroarene oxidation a reaction
temperature of 80-100 °C was found to be necessary for
the oxidation of acetophenones using the modified alka-
line nitrobenzene method.7 At the melting point of 60 °C
sodium perborate (NaBO3‚4H2O) decomposes, and it is
thus reasonable to assume that the decomposition tem-
perature of NaBO3‚4H2O and the required oxidation
temperature result in a mismatch for this oxidative cycle.
ESR studies performed several years ago by Russell
and co-workers17-19 have shown that nitrosobenzene 15
and hydroxylaniline 16 (Scheme 3) react in basic medium
to form two nitrosobenzene radical anions 17. Moreover
these ESR studies also revealed that if the reaction was
performed in the presence of molecular oxygen, the
This knowledge of the behavior of nitrosobenzene and
hydroxylaniline derivatives has been utilized to construct
a catalytic cycle using 1,3-dinitrobenzene as catalyst with
molecular oxygen as the terminal oxidant. The catalytic
cycle is outlined in Scheme 4.
Pathway (a) shows the initiation step that also oper-
ates as the oxidation step for the catalytic cycle. Step (a)
is constituted by the mechanism that operates in the
stoichiometric oxidation process7 as proposed in Scheme
1. The basic conditions originally present for the produc-
tion of the enolate anion of the acetophenone now serve
a dual purpose, since basic conditions are crucial for the
regeneration of the reduced nitroaryl as shown by Russell
and co-workers17 (Scheme 3). In the present oxidation
process the products 7 and 10 formed during the oxida-
tion reaction are converted into the reactive species 21
(nitrosoradical anion) following pathway (b). In the
presence of molecular oxygen the reoxidation proceeds
to provide the coupling product 23, which decomposes to
form the nitrobenzene radical anion 24 (reaction path-
ways (d) and (e)). The radical anion 24 is as the ultimate
step (f) of the catalytic cycle oxidized in a SET reaction
to afford 1,3-dinitrobenzene. This can be explained by the
fact that nitrosobenzene is a better electron acceptor than
nitrobenzene, as has been demonstrated by ESR experi-
ments.17 A termination reaction arises when two 1-nitro-
3-nitrosobenzene radical anions 21 react following path-
way (c) to give N,N′-bis(3-nitrophenyl)diazene N-oxide 22.
(14) McKillop, A.; Tarbin, J . A. Tetrahedron 1987, 43, 1753.
(15) McKillop, A.; Kemp, D. Tetrahedron 1989, 45, 3299.
(16) McKillop, A.; Sanderson, W. R. Tetrahedron 1995, 51, 6145.
(17) Russell, G. A.; Geels, E. J .; Smentowski, F. J .; Chang, K.-Y.;
Reynolds, J .; Kaupp G. J . Am. Chem. Soc. 1967, 89, 3821.
(18) Russell, G. A.; Geels, E. J . J . Am. Chem. Soc. 1965, 87, 122.
(19) Geels, E. J .; Konaka, R.; Russell, G. A. Chem. Commun. 1965,
13.
J . Org. Chem, Vol. 67, No. 21, 2002 7495

Evaluation Only. Created with Aspose.PDF. Copyright 2002-2021 Aspose Pty Ltd.
Bjørsvik et al.
TABLE 1. Exp er im en ts P er for m ed for Exp lor a tion a n d
Op tim iza tion of th e 1,3-Din itr oben zen e-Ca ta lyzed
Aer obic Oxid a tion P r ocess
The introduction of tert-butyl alcohol as solvent and
potassium tert-butoxide as base allowed the oxidation
method to operate successfully to afford desired carbox-
ylic acid. We believe that the tert-butoxide anion operates
only as a base without involvement in the SET process,
which is in agreement with earlier reports.12,20-22 Except
from the substrate and the corresponding carboxylic acid,
no other compounds were observed in the reaction
mixture. Experiments that demonstrated low selectivity
(Tables 1-3) contained only tars in addition to desired
carboxylic acid. In general, absence of 1,3-dinitrobenzene
disfavored the desired oxidation reaction and resulted in
lowered yield and selectivity.
An experiment performed without bubbling oxygen
(entry 27, Table 1) but with 5 mol % of 1,3-dinitobenzene
present afforded less than 10% of benzoic acid and thus
supported our belief that the reaction could be ascribed
to the stoichiometric oxidation with 1,3-dinitrobenzene
as oxidant, Scheme 1.
When the oxidation process was performed without the
mediator 1,3-dinitrobenzene present (entry 26, Table 1),
a surprisingly high yield of benzoic acid was determined,
namely, 63.7% with a conversion of 96.2%, which however
corresponds to a selectivity of only 66.2%. The base-
catalyzed autoxidation indicated as pathway (g) of Scheme
4 thus seems to be a substantial contributor to the overall
picture. Such a base-catalyzed aerobic oxidation of methyl
aryl ketones was described in the old literature.23 Later
reports by Wallace et al.24 and Petricˇ et al.25 describe
similar processes. These processes suffer however from
serious “environmental drawbacks” because the two
processes require the rather environmentally unfriendly
solvents hexamethylphosphoramide (HMPA) and dimeth-
ylformamide (DMF), respectively. Moreover the process
using HMPA as reaction medium is also burdened by a
very long reaction time.
Depending on the experimental conditions large varia-
tions of the obtained yield and selectivity were observed
for what we initially assumed to be the composite process
outlined in Scheme 4. These variations were explored
systematically when a series of experiments comprising
entries 1-26 of Table 1 were performed. The experiments
of Table 1 constitute a 21 × 32 factorial design.26 The two
experimental variables x1, quantity of t-BuOK, and x2,
reaction temperature, were initially explored on three
levels, while a third variable x3, quantity of the catalyst
1,3-dinitrobenzene, was explored at two experimental
levels. An empirical model shown in eq 1 relates the yield
reaction conditionsa
t-BuOK 1,3-Ph(NO3)2 conv selectivity yield
entry [mmol] [°C] O2
measured responsesb
T
[mmol]
0.12
[%]
[%]
[%]
1
2
3
4
5
6
7
8
2.4
2.4
4.8
4.8
7.2
7.2
2.4
2.4
4.8
4.8
7.2
7.2
2.4
2.4
4.8
4.8
7.2
7.2
8.0
8.0
7.2
7.2
7.6
7.6
8.0
8.0
7.2
15.0
20
20
20
20
20
20
50
50
50
50
50
50
80
80
80
80
80
80
90
90
90
90
85
85
80
80
80
80
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
-
+
55.8
44.7
67.0
52.0
73.0
69.2
50.0
64.5
74.6
66.8
76.6
85.0
75.8
70.5
72.9
88.3
92.5
91.7
75.8
94.2
87.9
95.6
91.4
94.6
95.8
96.2
13.3
23.1
25.9
59.6
51.7
57.7
44.6
71.6
53.8
71.5
71.2
92.9
53.9
44.0
48.1
84.0
39.2
87.4
63.6
98.9
53.5
86.0
66.1
80.0
59.9
87.8
66.2
60.0
13.0
11.6
39.9
26.9
42.3
31.0
35.8
34.7
53.5
47.6
71.3
45.8
34.0
33.9
61.2
34.6
80.8
58.2
75.2
50.6
75.6
63.2
72.9
56.8
84.2
63.7
<9
0.12
0.12
0.12
0.12
0.12
0.12
0.12
0.12
0.12
0.12
0.12
0.12
9
10
11
12
13
14
15
16
17
18
19c
20
21c
22
23c
24
25c
26
27d
28e
0.12
0.12
a
Procedure: Acetophenone (0.288 g 2.4 mmol) was dissolved
in t-BuOH (12 mL), followed by addition of t-BuOK and 1,3-
dinitrobenzene (0.12 mmol) with oxygen bubbling through the
reaction mixture. The reaction mixture was heated with stirring
b
for a period of 5 h. Conversion selectivity and yield are based on
isolated products and recovered substrates. Isolated substances
are corrected for impurities by GC analysis. Moreover, isolated
substances were also analyzed on GC-MS and NMR for identifica-
tion. c Experiment based on model for searching optimized condi-
d
tions. Experiment performed without bubbling oxygen but under
nitrogen atmosphere. The isolated product was impure. e NaOH
was used as base. NaOH was added as 30% (7.5 M) aqueous
solution.
It is reasonable to believe that the reaction following
pathway (c) may be restrained by increased oxygen
pressure that would favor oxidation over dimerization of
the 1-nitro-3-nitrosobenzene radical anions 21.
The new oxidation method described above is per-
formed in the absence of water and with a strong base
present for the formation of enolate anion from the
methyl aryl ketone and for the regeneration of reduction
products 7, 10, 11, etc. derived from the mediator 3. 1,3-
Dinitrobenzene is used in catalytic quantities (initially
we used 5 mol %) with molecular oxygen as the terminal
oxidant. When water is used as solvent as in the
stoichiometric process,7 the oxidation cycle fails (entry
28, Table 1).
(20) Guthrie, R. E.; Weisman, G. R.; Burdon, L. G. J . Am. Chem.
Soc. 1974, 96, 6955.
(21) Guthrie, R. E.; Weisman, G. R. J . Am. Chem. Soc. 1974, 96,
6962.
(22) Guthrie, R. E.; Pendygraft, G. W.; Young, A. T. J . Am. Chem.
Soc. 1976, 98, 5877.
(23) Miller, W.; Rohde, A. Ber. Dtsch. Chem. Ges. 1892, 25, 2095.
(24) Wallace, T. J .; Pobiner, H.; Schriesheim, A. J . Org. Chem. 1965,
30, 3768.
(25) Zabjek, A.; Petric, A. Tetrahedron Lett 1999, 40, 6077.
(26) Factorial design is a type of statistical experimental design
It has been shown by ESR studies13 that proton
interaction between the aromatic anion and the solvent
(i.e., hydrogen bonding) or a polarizing effect of the
oriented ionic environment as existing in the polar
solvents may influence change in the nitro group elec-
tronic distribution. Water has a higher dielectric constant
versus t-BuOH or other alcohols, so the influence is
stronger and can inhibit the redox process.
constructed for gaining maximum information from
a minimum
number of experiments. The regression models obtained from such
experimental plans comprise terms that explain the interaction
between the different experimental variables. A full account of such
methodology is give in (a) Box, G. E. P.; Hunter, W. G.; Hunter, J . S.
Statistics for Experimenters, An Introduction to Design Data Analysis
and Model Building; Wiley: New York, 1978. (b) Box, G. E. P.; Draper,
N. R. Empirical Model-Building and Response Surfaces; Wiley: New
York, 1987.
7496 J . Org. Chem., Vol. 67, No. 21, 2002